Statement:
Assume $V$ is a vector space over a field $K$. Let $\phi: V \rightarrow V$ be a linear map.
Then $\lambda \in K$ is an eigenvalue of $\phi$ iff $\lambda$ is a root of the characteristic polynomial $P$.
Proof:
Let $M$ be the matrix of $\phi$ and let $\lambda \in K$ be arbitrary.
We have $P(\lambda) = \det(\lambda E_n - M) = 0$ iff $\lambda Id_V - \phi$ is not bijective (and not injective). This follows from the theorems ...
This is equivalent to
$Eig_\lambda(\phi) = kern(\lambda Id_V - \phi) \neq 0$, which means that the eigenspace of $\lambda$ is not the nullspace. Therefore, $\lambda$ must be an eigenvalue of $\phi$.
The theorems he mentioned are the following:
Let $M$ be a $n \times n$ matrix. Then the following statements are equivalent:
- $det M \neq 0$.
- The rows of $M$ are linear indepentent.
- $M$ is invertible.
- $rank(M) = n$.
and
Assume $V$ and $W$ are vector spaces over $K$ with $dim(V) = n$ and $dim(W) = m$. Let $\phi: V \rightarrow W$ be a linear map that can be denoted by the matrix $M$. Then the following statements are true:
- $\phi$ is injective iff the columns of the matrix are linear independent.
- $\phi$ is surjective iff the columns of the matrix generate $K^m$.
- If $m = n$, then $\phi$ is bijective iff the columns of the matrix are a base of $K^m$, and this holds iff $M$ is invertible.
Questions
My question is: How do we know that $\phi$ is not injective given these two theorems and the argumentation above?